Background: Brain metastases affect one third of adult cancer patients. Stereotactic radiosurgery (SRS) is standard of care for patients with limited brain metastases. Yet most patients will experience post-treatment cognitive decline given the potential for high doses to eloquent white matter and the hippocampus. Objective/Hypothesis: Our team has developed innovative, robust imaging methods and automated segmentation techniques to identify critical white-matter tracts and the hippocampus using advanced diffusion tensor imaging (DTI) and volumetric imaging. These novel imaging techniques also allow us to directly and non-invasively measure microstructural changes after RT to critical brain structures in vivo. We will use these advanced imaging technologies in a prospective trial of cognitive-sparing brain SRS for brain metastases patients. Specific Aims: 1: To evaluate whether relative sparing of eloquent white matter tracts (critical for memory, language, attention, and executive functioning) and hippocampi from high doses during brain SRS results in improved 3-month post-SRS cognitive performance relative to historical controls in patients with 1 to 3 brain metastases. 2: To measure longitudinal trends in white matter damage (using DTI) and hippocampal atrophy (using volumetric change) among patients receiving cognitive-sparing brain SRS and correlate these imaging biomarkers with domain-specific cognitive outcomes. Study Design: We will prospectively enroll 60 adult patients with 1-3 brain metastases who are eligible for brain SRS and MRI. Patients will undergo MRI with DTI and 3D volumetric imaging at baseline (pre-SRS) and 1 month, 3 months, and 6 months afterwards. White matter and hippocampal segmentation will be performed and critical regions integrated into cognitive-sparing brain SRS planning with automated knowledge-based optimization. Cognitive-sparing dose constraints are derived from previous data. A well-established, validated battery of neurocognitive tests will be performed at baseline and 3 months post-SRS. Cognitive deterioration rate will be compared between the current trial and historical controls and linear regression used to analyze patient, tumor, and treatment related predictors of cognitive decline. Statistical modeling will be used to analyze changes in imaging biomarkers as a function of time and radiation dose, and these changes will be tested for association with domain-specific cognitive tests. Spatial sensitivity to RT dose across white matter tracts will be analyzed.